The Agent Skills Directory

[SAFE]: No malicious patterns detected. The skill implements standard bookmarking functionality, allowing users to save URLs and manage associated tags.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it processes untrusted user data (URLs).
Ingestion points: User-provided URLs and tag names entering the agent context via chat interaction.
Boundary markers: None explicitly defined in the skill instructions to separate the URL data from the agent's instructions.
Capability inventory: Limited to specialized tools for bookmarking: get_bookmark_tags, save_bookmark, and add_bookmark_tag. No capabilities for arbitrary command execution, network requests to unknown domains, or sensitive file access are present.
Sanitization: The skill instructions do not specify sanitization or validation; it is assumed that the underlying tool implementations (save_bookmark) handle the provided strings safely.

url-bookmark